The overall objective of this project is to study LAGLIDADG homing endonuclease (LHE)-based hematopoietic stem cell (HSC) gene repair strategies in a clinically relevant large animal model, the dog. We have used the dog to develop improved HSC gene transfer strategies and to evaluate different gene transfer systems. A particular advantage of the dog model is the availability of disease models such as X-linked severe combined immunodeficiency (XSCID) and pyruvate kinase (PK) deficiency. The development of leukemia in three children with XSCID after HSC gene therapy with a gammaretroviral vector has demonstrated the risks of retrovirus-mediated gene therapy. Furthermore, our preliminary data in the dog demonstrate that all three major integrating retroviral vectors, i.e. gamma, lenti, and foamy .virus vectors carry a substantial risk of insertional mutagenesis. Thus, the development of site-specific gene replacement/gene repair strategies with reduced risks of insertional mutagenesis has become a crucially important goal for the development of safe gene therapy/repair approaches in patients with genetic disorders. To address the feasibility and safety of site-specific double strand break-induced gene targeting driven by LHEs, we will study and optimize conditions for use of self inactivating non-integrating lentiviral (NIL) vectors to drive LHE-mediated "genomic marking" of dog CD34+ repopulating cells. In Aim 1 we will target a cleavage site for an existing l-Scel variant enzyme in a non-transcribed region of the genome, while in Aim 2 we will generate dogs with an l-Scel cleavage site embedded within a fluorescent reporter. As part of Components 2-5, new LHEs capable of cleaving alternative sites will be generated; reporters with cleavage sites for these LHEs will then be developed, and the new LHEs performance benchmarked against l-Scel. In aim 3 we will assess the feasibility of a site-specific knock-in of a selectable marker gene such as the P140K mutant of methylguanine methyltransferase (MGMT) and then use nonmyeloablative conditioning regimens in combination with in vivo selection, and in aim 4 we will evaluate the optimized conditions with LHEs created by Components 2-5 in two canine disease models, the XSCID and the PK models. We believe these studies will allow us to advance the use of LHEs in a clinically relevant large animal model.